"Well-known science fiction author Walter Koenig ("Buck Alice and the Actor Robot"), who also had a bit part in some obscure 1960s space show, has died".

While it doesn't seem likely last minute 2 hour extension could have changed the results this was not fair to voters given the extensions were not implemented uniformly across all polling locations. Some locations had already closed up and shut off their machines.

Sorry I meant moat as in the phrase "there is no moat" from the classic google memo of a few years ago.

https://newsletter.semianalysi...

What I've heard is a market cap well north of a trillion dollars is expected.

It's not just the losing money aspect by itself that is concerning. People could overlook this when they foresee a credible prospect for turn around. It is losing money combined with being a commodity + rapid obsolescence that seems strange to me. There does not as far as I can see seem to be much in the way of being reasonably able to predicting any sort of future for Anthropic.

Are people really going to put trillions of dollars into a company that has never made a profit and continues losing billions of dollars annually? A company whose current technology has no mote and will be obsoleted in mere months?

I've never been good at understanding markets. During the .com era found myself as bewildered by the crazy math going around as I am today. At least my strategic reserves of popcorn were put to good use.

I think it best to let the hardware deal with error correction. It is way better positioned in the stack to judge and handle these sorts of failures at least cost. At best all you need are simple checksums to alert the operator to hardware failure.

This is strange. SSDs are supposed to have scrubbers that periodically scan for this and rewrite leaky blocks internally.

I don't mind being shit on but Jesus at least put some effort into it.

Or if you're really worried about these types of errors you could just use one of the error-recovery-targeted RAID levels and not have to worry about the massive performance hit of RS-encoding and -decoding each data block read...

Questioning something and pushing back with your own counter arguments is called arguing.

arguing: "to give reasons for or against something"
https://www.merriam-webster.co...

It doesn't matter how nice you are or how you elect to characterize the activity itself ... these types of interactions /w LLMs are a waste of time.

All I said was arguing with LLMs is a waste of time and you immediately responded with derisive commentary and continued to do so.

It is a well known fact LLMs have no meta-cognition. There is no internal mechanism that allows for it.

"We find no evidence for significant verbalised individuated metacognition in any tested domain"
https://arxiv.org/pdf/2605.242...
https://www.nature.com/article...

LOL leave it to the one who doesn't know what the word "arguing" means to judge the intelligence of others.

You: "If you question something it says push back with your own counter-arguments."

Me: "It is hard to imagine a bigger waste of time than arguing with an LLM."

You: "The fact that you think having a conversation and an argument are the same thing tells us everything we need to know about you."

Yes, I am absolutely projecting the words you yourself just used back in your face.

Did AI help you write this or did you come up with it all by yourself?

Your own words "If you question something it says push back with your own counter-arguments." this is called arguing. You yourself use the word "arguments". Do you think if you converse your "counter-arguments" with it nicely and diplomatically so as to not hurt its feelings the experience will be more useful?

LLMs have no meta-cognition of any kind whatsoever. They don't know what they know. They don't know what they don't know. They don't even know what know is.

It is hard to imagine a bigger waste of time than arguing with an LLM.

I'm not sure using revenue as a metric is all that relevant. It is possible for AI to radically increase or decrease labor productivity and yet a corporation still take in the same revenue.

Same problem with regards to profit albeit more relevant than revenue. It is possible for labor productivity to increase yet not translate into profitability due to other factors... competitive pressures (e.g. competitors making the same investments) and contravening costs.

I think the bottom line of whether AI is in aggregate helpful should be measured in terms of labor productivity. In IT when it comes down to it you end up playing shell games with value. While you can give people capabilities and insights they might not otherwise have reliance and changes within IT and personal skill sets erodes other capabilities that may well have offered similar streams of value. People generally seem to be terrible at reasoning about opportunity costs.

Decades ago I loved watching our accounting people navigate TUIs from their keyboards.. they could do anything in seconds.. it went so fast the screen was a blur... Did the fancy slow as shit web version of what replaced the ancient software increase productivity? Somehow I doubt it. Just look around at basic groupware / messaging... basic functionality to support productivity seems to be at best an afterthought. What matters most is how shiny and fancy UX is and ease of use. Not actual responsiveness or productivity. How anything can be slow as shit given the hardware available today sometimes I have trouble believing.

Or maybe diminishing returns associated with new technology means more than you think it does.

The computer I am typing this on is alien technology compared to what was available when I started. I remember when major national backbones had less capacity than the fiber CE in our home. While the underlying technology is amazing there has always been this layer of value taking, sleaze and laziness that has mostly nullified extraction of even incremental value amongst the masses from improved technology.

The transformer based AI from a technology perspective is something I never imagined I would live to see. It is by all rights fucking insane compared to what existed before it. While useful and all I'm not even sure on balance this actually translates to a net benefit. What do you even do with something that is completely untrustworthy and often spews nonsense? What of the considerable collateral damage the technology is enabling?

AI to me is kind of like a self driving car... it would be one thing if the car actually drove itself and didn't require a licensed driver to sit behind the wheel and dedicate their full undivided attention to making sure the car doesn't do something stupid... yet no such thing at present exists.

Too much technology is squandered on race to the bottom path of least resistance garbage the tech industry keeps shitting out having adopted business models indistinguishable from the malware vendors of old.

Probably not, but they will eventually target you: "flout international law or funnel coveted classified contracts to political allies" will also include "target anyone not showing puppy-like devotion to the Dear Leader".